Keywords

1 Do We Truly ‘Share’ the Moral Law? Thompson’s Challenge to Kant

Kant ties the notion of a ‘duty towards someone’ to the idea of ‘necessitating’ or ‘constraining’ one another under the moral law (MM 6:442.10; see also MM 6:241.15–17). This is his stated reason for restricting moral status to persons, excluding animals (MM 6:442.11–12). Although Kant never makes explicit what this ‘constraining’ activity consists in, it is clear from the context that only individuals who mutually share the moral law can constrain each other in this way (MM 6:442.11–12; Wood 1998, 189; Korsgaard 2018, 123ff.; Ripstein and Tenenbaum 2020, 147). So when it comes to the conditions of the possibility of duties towards others, the animal advocate has a bone to pick with Kant.

In this chapter I propose a way to amend Kant’s conception of moral duties so that moral duties can be directed towards animals. For my argument, I enlist some outside help from critics who object that moral agents, as Kant conceives of them, cannot genuinely share the moral law in the first place. Thompson (2004) has argued that since the moral law is autonomous, individual Kantian agents stand under separate moral laws with the same content each, but not under one and the same, shared moral law. The upshot for Thompson is that Kant quite simply fails to account for the directionality of duties. The usual Kantian reaction has been to come up with an account of how the moral law is shared between Kantian persons after all (Fanselow 2008; Darwall 2009; Palatnik 2018). For reasons I will explain in this chapter, this is not my preferred response. I argue that we can, and should, account for directionality in an entirely different way and that Kant provides all the resources we need.

In a nutshell, I suggest we should account for directionality in terms of the content of our duties. Duties towards others are just duties that are about others in a certain special way. Likewise, duties towards self are just duties that are about ourselves in a certain special way. But of course, duties can be about others even if those others do not share the moral law. So with my conception of directionality, I remove any elements of what today’s ethicists sometimes call ‘second-personal ethics’ from Kant’s system. What remains is a version of Kant’s ethics that doubles down on the “fundamentally and radically first-personal” features it already has (Timmermann 2014, 131; see also Sensen 2011, 120f.). I suggest it is all the better for it, certainly from the perspective of an animal ethicist. On the ‘first-personal’ conception of directionality, there can very well be duties towards animals, and they do not categorically differ in form or ground from our duties towards other human beings.

To approach the notion of directionality, consider that we have moral duties to or towards specific subjects. For instance, we seem to owe it to another person not to inflict pain on them. The other person is not just a location where we can fulfil a blanket duty not to inflict pain. It is no coincidence that only the affected person can consent to having pain inflicted on them, for instance. Neither is it a coincidence that if we have inflicted pain on someone, we must apologise specifically to them, and only they can forgive us. These are just some of the ways in which the directionality of duties matters.

Kant would affirm all of this. As we have seen in Chap. 2, Kant takes us to have duties towards others and duties towards self. No duty is directed at no one. Kant furthermore pays close attention to the distinction between a duty towards someone and a duty merely regarding them (see MM 6:442.04–06). However, despite this eagerness to incorporate directionality into his system of moral philosophy, according to Thompson (2004), Kant has trouble accounting for it on his own terms.

Before explaining the problem in more detail, let me introduce some of the special vocabulary used in this debate. Thompson draws a distinction between ‘monadic’ and ‘bipolar’ normative judgements. Monadic normative judgements are judgements such as ‘X did wrong in doing A’, ‘X has a duty to do A’, and ‘X has a right to do A’ (Thompson 2004, 338). These judgements make an explicit reference only to one subject. Contrast this with judgements such as ‘X wronged Y by doing A’, ‘X has a duty to Y to do A’, and ‘X has a right against Y that he do A’ (ibid., 335). These make reference to two subjects and are hence called ‘bipolar’.Footnote 1 For Thompson, bipolar normative judgements correspond to bipolar normative relations or what he calls ‘dikaiological orders’ under which subjects stand (Thompson 2004, 352). All subjects who can be part of the same bipolar normative relation are called ‘persons’ relative to that relation (ibid.). For instance, all subjects who can stand in the relation of ‘morally owing someone something’ are moral persons. Similarly, all subjects who can enter into legal relations under Roman law are legal persons under the system of Roman law. Finally, the person who bears a duty is called the ‘obligor’, while the person towards whom the duty is directed is called the ‘obligee’ (Darwall 2012, 333).

Among other things, it matters how we account for the directionality or ‘bipolarity’ of moral duties because our account will determine the scope of subjects towards whom these duties can be directed.Footnote 2 Thompson’s central idea here is that we can only account for bipolarity in terms of standing under a shared law. This can be a natural idea if we consider contract law: Two individuals can only enter into a binding contract with each other if they both stand under the same system of contract law.Footnote 3 And similarly, if there are to be binding bipolar moral relations, it seems they can only obtain between individuals who stand under a shared system of moral law. ‘Thompson’s puzzle’, as it is sometimes called, is the challenge to account for the bipolarity of moral duties in terms of a shared moral law without dubious assumptions. Thompson claims that Kant, along with Hume and Aristotle, is unable to account for bipolar moral relations in terms of a shared moral law, or at any rate that he would have to rely on “alarming metaphysical commitments” to do so (Thompson 2004, 379). It is helpful to consider this difficulty in more detail because it sheds light on what it means to ‘share’ the moral law, and why this sharing is so important for interpersonal duties.

Why would Kant fail to account for bipolar moral relations in terms of a shared moral law? Thompson’s argument is best illustrated by his own example (Thompson 2004, 361f.): Say the Lombard people have their own system of laws under which all Lombards, and only the Lombards, are legal persons. And say that on the other side of an Alpine mountain range, by sheer coincidence, a people has independently developed which is indistinguishable from the Lombards, down to the language and the legal system. Call these people the Schlombards (though they, of course, call themselves the ‘Lombards’). Thompson—convincingly, to my mind—argues that Lombards and Schlombards do not live by the very same law, given that their cultures developed in complete isolation from each other. If a Lombard and a Schlombard should happen to meet on the mountain range and attempt to enter into a binding contract with each other, they would fail. They could not enter a binding contract under either Lombard or Schlombard law, since each is not a legal person under the legal system of the other.

The objection to Kant is that his moral philosophy puts moral agents in the position of Lombards and Schlombards (Thompson 2004, 382). For Kant, the moral law is only binding because it is autonomous. It is a self-imposed law. It can only be autonomous, however, because it is also autochthonous. That is, it is only a self-imposed law because it originates in the individual rational being (see Timmermann 2007, 114f.).Footnote 4 Therefore, even though all moral agents individually give moral laws to themselves that have the same content each, they do not stand under the very same law. Their moral laws are distinct in origin and authority. Since they do not stand under the very same moral law, they cannot enter bipolar moral relations under that law, argues Thompson. So duties ‘towards’ other people are impossible for Kant, if Thompson is right. There are only blanket duties, and other people are mere locations where we can fulfil them.

Thompson acknowledges that there is one way in which Kant can possibly account for bipolar moral relations. Kant might account for bipolarity by claiming that the moral law does after all have a common origin in all moral agents, namely in pure practical reason, which is not individuated (Thompson 2004, 382). On the one hand, however, this would put a lot of argumentative weight on the claim that pure practical reason is the singular or unified cause of the moral law, and Thompson takes it that this would be an “alarming metaphysical commitment” (Thompson 2004, 379). Be that as it may, the analogy to Lombards and Schlombards would suggest that in order to be the very same law on Thompson’s terms, the moral law would need a common natural origin in all moral agents, and it is at least not obvious that a common nonnatural origin in pure practical reason will do the trick. It seems, therefore, that Kant indeed faces a difficulty in accounting for bipolarity in terms of a shared moral law.

Kantians have come up with several answers. The most enthusiastic response to Thompson comes from Darwall (2009), who agrees that an autonomous law alone cannot account for bipolar moral relations (Darwall 2009, 138). Accordingly, Darwall advises that a Kantian account of the moral law should start not from the notion of an autonomous law alone, but from the notion of second-personal accountability (Darwall 2009, 153). This response to Thompson’s challenge obviously comes at a steep price. To agree with Darwall, Kantians would have to sacrifice the view that the moral law is autonomous. But that is to sacrifice an absolutely central element of Kant’s moral philosophy.

Fanselow (2008) proposes another response: Moral agents are not merely bound by the moral law in general, but also by specific duties. These duties follow from the Categorical Imperative only with the addition of facts about the world. For instance, the duty to hold one’s promises derives—on Fanselow’s viewFootnote 5—from the Formula of Universal Law (FUL) with the addition of certain facts about promising and human psychology (Fanselow 2008, 96). The idea is that the duty not to make false promises follows from the FUL only if we add the fact that in a world of false promisers, human beings would not take promises seriously. Since moral agents are all part of the natural world, they share its facts. So even if the moral law is not truly shared among Kantian moral agents, their materially specific duties are based on shared facts. According to Fanselow, this is enough of a shared origin to enable bipolar moral relations at the level of specific duties, and that is what matters.

The main problem with Fanselow’s response is that when Thompson asks for a shared origin that accounts for the bindingness of a bipolar moral relation, not just any shared origin will do. By way of analogy, the specific contract into which the Lombard and the Schlombard attempt to enter has a shared natural origin in some respects. The Lombard and the Schlombard have met on the same mountain pass on the same day, have written their contract on the same piece of parchment, and so on. But none of this makes it the case that they indeed stand in a binding legal relation. For that, they would need a common system of contract law. Hence, what would be needed is a common origin of the system of law under which the contract is binding, not just a common origin of the contract itself. And since the Lombard and the Schlombard do not stand under the same system of contract law, their contract lacks the right kind of common natural origin. The same holds for Kantian moral agents. Even if they derive their duties from the Categorical Imperative using the same facts, the moral law which makes these duties binding still lacks a common natural origin, and so there is no common normative ground that can account for the bindingness of the Kantian bipolar moral relation. Hence, Fanselow’s response is not enough to resolve Thompson’s puzzle.

According to another Kantian response (Palatnik 2018), Kantian moral agents usually represent each other as beings with a will. On Palatnik’s account, Kantian agents recognise each other in the world of appearances on the basis of “regularities or patterns characteristic of intentional activity” (Palatnik 2018, 294). It thus comes naturally to agents to recognise and represent others as rational beings. Representing others as rational beings however necessarily involves representing them as being bound by the moral law—the same moral law under which the agent herself stands (Palatnik 2018, 295). Thus, “we cannot represent (and recognise) anyone as a practical reasoner, as a person in relation to ourselves, unless we can see him as sharing our basic framework of practical thought, as falling under the principles of pure practical reason and fundamental concepts that follow from these principles” (Palatnik 2018, 300). Therefore, Palatnik’s suggestion is that for Kant, a moral agent naturally recognises and represents others as beings bound by the same moral law by which she considers herself to be bound.

The claim that Kantian agents can recognise rational—that is, autonomous—beings by natural regularities or patterns might be hard for strict Kantians to accept, but this need not concern us here.Footnote 6 What matters is what options are open to Kantians, not whether Kant himself would endorse them. However, Palatnik’s suggestion faces a great challenge in solving Thompson’s puzzle: It can only account for the bipolar moral relations in which moral agents take themselves to stand. It explains why we consider ourselves to be bound by bipolar moral relations. But Thompson’s puzzle asks for an account of the bipolar relations in which moral agents actually stand. For example, the Lombard and the Schlombard represent each other as persons under the same law, but they still fail to enter a binding contract because they do not actually stand under the same law. Thus, Palatnik’s response is not sufficient to resolve Thompson’s puzzle either.

In summary, previous Kantian responses to Thompson’s puzzle either face great difficulties, as I argued in the case with Fanselow and Palatnik, or come at a very high cost, as I argued in the case with Darwall. Of course, one could continue the conversation. My point is not that no Kantian solution to Thompson’s puzzle is possible. What I hope to have shown, however, is that even the most competent Kantian philosophers have had a surprisingly hard time explaining how a shared moral law is supposed to account for the directionality of Kantian moral duties. This is an area where Kantian ethics stands in dire need of clarification. And this makes it an attractive area for modification.

Consider further that the various responses all strengthen Kant’s denial of duties towards animals. On Darwall’s account, bipolarity arises from practices or attitudes of holding each other second-personally accountable, in which animals do not participate. On Fanselow’s account, bipolarity arises from deriving duties from the Categorical Imperative using facts about the shared natural world, another activity in which animals do not partake. And on Palatnik’s account, bipolarity arises from an agent’s supposition that the other is bound by the same moral law as herself, which once again is not true for animals. If Kantians want to be able to account for duties towards animals, they should look for a different response to Thompson’s puzzle.

It is no coincidence that previous responses to Thompson’s puzzle end up excluding animals all the more vehemently. After all, they share with Thompson a crucial assumption: that we must account for bipolar moral relations in terms of a shared moral law. So long as we come up with ways for Kantian ethics to accommodate a shared moral law, the resulting moral philosophy will inevitably exclude animals from moral concern. I therefore suggest a more radical, but also more elegant response to Thompson’s puzzle: Kantians should refuse to solve it. They should deny altogether that bipolarity, in Thompson’s sense of the word, is a feature of the moral landscape. And where there is nothing to be accounted for, there is no puzzle. This is a more robust Kantian response to Thompson, and at the same time it removes a major obstacle for the inclusion of animals in Kantian moral concern.

2 First-Personal Versus Second-Personal Accounts of ‘Directionality’

To reiterate, Kant himself clearly thinks that only rational beings who mutually share the moral law could have duties ‘towards’ each other. The underlying reasoning is that only beings who share the moral law can ‘necessitate’ or ‘constrain’ each other (MM 6:442.10; see Wood 1998, 189; Korsgaard 2018, 123ff.). It is noteworthy that Kant says the same thing about the juridical domain—animals without legal duties cannot have legal rights because they cannot obligate others, he argues very briefly (MM 6:241.15–17). Thus, the suspicion is not too far-fetched that Kant, like Thompson, draws on an analogy between morality and legal systems when he ties bipolar relations to the sharing of a law. However, I will now suggest that Kant can also inspire a rather different understanding of bipolarity or directionality. This alternative understanding does not rely on the notion of a shared moral law. So we can save Kant with the resources he provides.

To start with, consider some differences between Kant and Thompson when it comes to moral bipolarity. For Thompson, monadic duties form a distinct type from bipolar duties. For instance, according to Thompson there is a monadic duty not to lie, over and above the bipolar duty towards the conversation partner not to lie (Thompson 2004, 339f.). This corresponds to the intuition that we can feel guilty simply for having lied, not only for having wronged someone by lying (Thompson 2004, 340). Kant would not draw the distinction in this way. For Kant, all duties are bipolar or ‘directed’ towards an obligating subject, and there is no such thing as a genuinely monadic or ‘undirected’ duty. Consider the structure of the Doctrine of Virtue: Here, Kant makes an exhaustive distinction between duties to self and duties to others (MM 6:413.07–08). There is no third option. What about Thompson’s example of feeling guilty simply for having lied? Kant would agree that a liar has more to feel guilty about than merely having wronged someone else, because we also have a duty towards self not to lie (MM 6:429.04–06). But he would flatly deny that there is an ‘undirected’ duty not to lie.

None of this is to say that Kant is obsessed with the bipolar nature of our duties. He is happy to omit the bipolar element from a duty’s description when it is irrelevant. For Kant, the monadic and the bipolar are simply modes of description. For Thompson, they are types of duties. This is a significant difference. In fact, when Kant and Thompson talk about ‘duties towards’, they may not be talking about the same thing. And so it is anything but obvious that bipolarity in Thompson’s sense must be a feature of the Kantian moral landscape at all. Kant is only committed to bipolarity in his own sense, not Thompson’s.

Another, though related difference between Thompson and Kant concerns ‘moral authority’. Think of authority as the capacity to create or eradicate duties for someone. Kant’s moral philosophy is ‘first-personal’ in that it derives duties from an autonomous moral law. He considers heteronomy of the will to be “the source of all spurious principles of morality” (G 4:441.02). In Kant’s view, then, only the rational being herself has moral authority. Of course, the rational being still has no capacity to create or eradicate duties at her whim. Still, duties only derive from the moral law the agent gives to herself. I will call this the ‘first-personal Kantian view’. Now, my radical suggestion is that first-personal Kantians should not bother arguing that the moral law is interpersonally ‘shared’ in any interesting sense. They should stick steadfastly to their claim that the moral law is autonomous and autochthonous, period. And so, for first-personal Kantians, there is no such thing as second-personal moral authority. Others do not impose the moral law on me, and they consequently cannot create or eradicate any of my duties.

Contrast this view with a ‘second-personal Kantianism’ along Darwall’s lines, which would derive duties from a moral law that is in turn grounded in second-personal accountability (Darwall 2009, 153). The idea is, in a nutshell, that the moral law only arises from acts of second-personal address such as making claims on each other, objecting to each other, holding others answerable, and so on. These acts can, but do not have to, consist in verbal utterances. They can also be reactive attitudes. For instance, we hold others accountable for their wrongdoing simply by resenting them. If the moral law derives from such acts of second-personal address, it is an essentially interpersonal law. An important upshot of this is that others have moral authority (‘second-personal authority’). They can put others under obligation or lift obligations off their shoulders.Footnote 7 For instance, by resenting me, another person puts me under obligation to apologise. By forgiving me, she lifts this obligation off my shoulders.

The view that others have moral authority helps second-personal Kantians account for the bipolarity of duties. We can simply say that duties are directed ‘towards’ the individual whose moral authority makes it binding.Footnote 8 Evidently, this option is not open to first-personal Kantians, who do not believe in this kind of moral authority. How, then, can they account for bipolarity?

Kantians typically insist that “directionality is determined by the source of constraint (your rational will), not by the nature of the resulting object of choice” (Ripstein and Tenenbaum 2020, 147). While this seems like a correct representation of Kant’s view, it is also a view Kantians can choose to reject. We can go the other way and argue that moral directionality is precisely a matter of what our duties are about. According to this view, a duty towards X is a duty that is about X in a certain way. Never mind who makes a duty binding—consider instead what duties ask us to do and why! Put in Kantian terms, bipolarity is a matter of the content or material of duty, that is, the end or action we ought to pursue (MM 6:398.5–20). It is not a matter of the form of duty, that is, that end or action’s character of moral prescription. Call this the content approach to bipolarity, as opposed to Thompson’s (and Darwall’s, Fanselow’s, Palatnik’s) form approach.

How exactly do duties have to be about X in order to be duties towards X? Interestingly enough, we can gather a version of the content approach from Kant himself. In particular, consider how Kant first introduces the distinction between duties to self and duties to others in the Doctrine of Virtue (MM 6:385f.). As we have seen in Chap. 4, Kant here moves from the notion of a Categorical Imperative to that of an ‘end that is also a duty’ (MM 6:385.31). That is, he moves from the law that prescribes an action unconditionally to the end of that prescribed action. He then draws the central distinction between the obligatory ends insofar as they pertain to our treatment of others and self (MM 6:385.31–386.14). This is, effectively, how he introduces the two duty-types. And with this conception in mind, we can flesh out the content approach to bipolarity:

The content approach to bipolarity

Any duty is a duty towards self iff it is about the promotion of moral self-perfection.

Any duty is a duty towards another person X iff it is about the promotion of X’s happiness.

For the sake of illustration, consider our duties towards self as Kant conceives of them. In particular, take the prohibition of prudential suicide. Our duty not to commit prudential suicide is a duty ‘towards self’ insofar as it is about our own moral perfection. We ought to keep our natural shape serviceable to morality, and of course being alive is a basic condition for duty-observance (see MM 6:421.10–14). Similarly, our duties towards others are those that are about the promotion of their happiness. Our duty of beneficence is a duty ‘towards’ others because it is about the promotion of their happiness. Kant’s treatment of the end that is also a duty therefore gives us all we need. The content approach to bipolarity suffices to account for the distinction between duties to self and others.

At this point, one might object that the content approach removes Kant’s distinction between duties towards someone and duties merely regarding them. After all, both duties are about the same individual. But this objection would miss an important point, namely that the content approach does not equate directionality with ‘being-about’ in general. For a duty to be a duty towards X, it must more specifically be about X’s happiness insofar as it is a part of the obligatory end of the happiness of others, or about X’s moral perfection insofar as it is a part of the obligatory end of one’s own moral perfection. Duties merely regarding X are duties that are about X, but not about X’s happiness or moral perfection as part of the obligatory ends.

Consider one of Kant’s own examples, the duty not to wantonly destroy beautiful plants (MM 6:443.02–04). This is a duty regarding the plant, not towards it. This is because the duty is about the plant, but not about the plant’s happiness or its moral perfection. This duty is all about the perfection of the moral agent as part of one of the obligatory ends. Hence, the content approach does not jettison the towards-regarding distinction.

Another objection could be that focusing on content in the proposed way blurs the line between duties towards self and others. One could rightly point out that for Kant, there exists at least one duty that is about our happiness but is a duty towards self: our indirect duty to secure a minimum of happiness for ourselves (G 4:399.03–07; MM 6:388.17–30). But this objection would be flawed too. Consider that this duty is about our own happiness only insofar as it forms a part of the obligatory end of our own moral perfection. We should not allow ourselves to become too unhappy, says Kant, because “want of satisfaction with one’s condition […] could easily become a great temptation to transgression of duty” (G 4:399.04–07). So even if this duty is about our happiness, it is only about our happiness as part of the obligatory end of our own moral perfection. In this way, the content approach upholds the line between duties to self and others.

However, there could be another, more philosophical objection to the content approach to bipolarity. Darwall explicitly argues against the view that duties towards others are just duties with a certain content:

We can easily imagine a society (Feinberg’s ‘Nowheresville’) in which it is thought morally wrong to step on others’ feet, unless, say, they desire or do not mind one’s doing so, but where the latter is not seen as a giving consent that can be understood only within a bipolar dikaiological order. So viewed, others’ will and preference would appear simply as features of the moral landscape that bear on moral obligations period. (Darwall 2012, 344f.)

The Nowheresville example illustrates that a duty towards X not to step on X’s foot is something more than merely a duty not to step on X’s foot. Rather, according to Darwall’s suggestion, duties towards X must be explained in terms of X’s second-personal authority, her capacity to obligate the agent under the moral law. By treating others “simply as features of the moral landscape”, we turn them into static factors of duty and overlook their subjecthood, or so Darwall worries. Once again, we run the risk of treating others merely as locations at which our pursuit of duties can take place (this worry also plagues Korsgaard 2018, 159). In the face of this objection, what can be said on behalf of the content approach?

It is quite intuitive that a duty towards X not to step on X’s foot is something more specific than merely a duty not to step on X’s foot. But in itself, that cuts no ice with the content approach. According to this approach, to say that a duty is directed ‘towards X’ does add something to the duty’s description: that it is a duty about the promotion of X’s happiness. In saying that a duty is directed ‘towards X’, we indicate that the route of reasoning to this duty passes through the obligatory end of others’ happiness. So the first-personal Kantian simply agrees with Darwall that the bipolar element adds something to a duty’s description. But it adds nothing that needs accounting for in second-personal terms.

What about Darwall’s more general worry that morality without second-personal elements leads to ‘Nowheresville’? Are others bound to be mere ‘features of the moral landscape’ on a first-personal view? Yes and no. First-personal Kantians deny that others can help bring about the moral landscape. The moral landscape exists because—and only because—the moral agent gives the moral law to herself. In this sense, others really are just features of the moral landscape. They register on a radar set up by rational beings from the first-person standpoint. But this does not imply that others cannot play a unique role in determining what actions are prescribed. Recall that for Kant, “what [others] may count as belonging to their happiness is left to their own judgement” (MM 6:388.08–09). The more we emphasise the role of another’s agency as a guide for our practical benevolence, the more active their involvement in our duties is. Therefore, we should not conceive of others as static features of the moral landscape. They are uniquely dynamic features. The first-personal Kantian view is sensitive to the fact that others are active subjects. What is denied is only that the subjecthood of others helps to create or eradicate any parts of the moral landscape.

3 Rejecting Thompson’s Challenge

Against this backdrop of Kantian conceptions, let me propose a novel response to Thompson’s puzzle. To reiterate, the puzzle is the challenge of accounting for moral bipolarity in terms of a shared moral law. Kantians have trouble meeting this challenge. I have now argued that Kant’s view of bipolarity differs from Thompson. In fact, the words ‘duty towards X’ may mean something quite different for the two.

First-personal Kantians cannot account for bipolarity in terms of a shared moral law. We must accept this. But that is not to admit defeat in the face of Thompson’s puzzle, to the contrary! Rather, first-personal Kantians can simply deny that bipolarity in Thompson’s sense exists. Our duties simply do not have any feature that needs accounting for in second-personal terms. To be sure, this is not a solution to Thompson’s puzzle, but a rejection of its presupposition. This rejection also avoids the difficulties of previous responses to the puzzle: controversial metaphysical commitments (Thompson’s Kant), the Lombard-Schlombard problem (Fanselow, Palatnik), and the exclusion of animals from direct moral consideration (all of the above, also Darwall).

It may seem that what I propose is easier said than done. After all, the conception of bipolarity we find in Thompson and Darwall is supposed to describe certain aspects of moral life. Moral life, the basic idea goes, consists not just of monadic duties. It consists of essentially interpersonal relations. We owe things to each other, object to each other, forgive each other, and so on. To use Thompson’s term, the moral landscape is full of ‘dikaiological orders’.

For second-personal Kantians, interpersonal relations uniquely connect obligor and obligee.Footnote 9 Part of this is that the obligee has a special standing vis à vis the obligor. Case in point: Only obligees have the power to consent (Darwall 2012, 345). And when they have been wronged, obligees have a special standing too: Only the obligee can forgive the wrongdoer, and only the obligee is owed apologies (Darwall 2012, 346). Naturally, Darwall would explain these features of moral life under reference to second-personal authority (Darwall 2012, 347). As first-personal Kantians who follow the content approach, are we to flatly deny that moral life has these features? Is everybody in a position to consent for anyone else? Can everybody forgive wrongdoers? Do wrongdoers have to apologise to everybody equally?

Fortunately, no such absurd claims follow. First-personal Kantians can account for ordinary practices of consenting, forgiving, and apologising, with different means. They can use what I will call the ‘deflate-and-deny tactic’:

The deflate-and-deny tactic

For any supposed exercise of second-personal authority, deflate it, as far as possible, to an exercise of the moral authority of the moral agent plus the capacity of others to determine the content of an agent’s duty. Deny the rest.

To give a preliminary example: Consider the ordinary practice of forgiving others for past wrongs. Darwall would consider forgiveness to be an act of release-from-duty. Among other things, to forgive someone is to release them from the duty to make amends. But we do not have the power to release others from any of their duties on a first-personal Kantian view. Either they have a duty to make amends that derives from the autonomously imposed moral law, or they do not. Us others have no say in the matter. What the first-personal Kantian can say, however, is that to verbally forgive is to provide information about the conditions of one’s happiness. Broadly, it is to signal that one is ‘over it’. The obligee verbally forgives the wrongdoer and thus makes it clear that no more amends are required for her happiness. This does not change anything about the wrongdoer’s duties, but it makes it clear that making further amends are not what their duties towards the obligee demand of them. This is the ‘deflate’ bit of the tactic. What the first-personal Kantian cannot account for is forgiveness as an act of literally lifting a duty off the obligor’s shoulders, and so she should simply deny that this is possible. We have no power to create or eliminate the duties of others. This is the ‘deny’ bit of the tactic.

Deflate-and-deny is a tactic because it can be applied in various contexts. No matter which specific examples adherents of second-personal authority bring up, deflate-and-deny can be applied. Of course, the tactic must prove its worth in particular applications, but it would be futile to attempt a comprehensive list of supposedly ‘second-personal’ aspects of moral life.Footnote 10 To second-personal Kantians, the moral landscape is saturated with second-personal authority. In Sect. 5.4, let me merely attempt to illustrate the proposed tactic by discussing Darwall’s most prominent, aforementioned examples: consent, forgiveness, and apologies. We will see that so much about the giving and obtaining of consent, about forgiving, and about apologising can be ‘deflated’—particularly to the determining of the content of another’s duties and informing them about it—that very little has to be ‘denied’. The first-personal Kantian thus does not have to reject these moral practices, just their explanation in second-personal terms.

4 Consent, Forgiveness, and Apologies Without Second-Personal Authority

Darwall makes it clear that he conceives of giving consent as “an exercise of a ‘normative power’, in this case, to release someone from a bipolar obligation he would otherwise have” (Darwall 2012, 345). Consent is thus, like forgiveness, an instance of release-from-duty. As we have already seen, we have no such power to create or eradicate duties for others on a first-personal Kantian view. Therefore, the deflate-and-deny tactic should be applied to release-from-duty, and by extension to consent.

A part of what Darwall means by ‘releasing’ someone from their duty can be deflated to powers the first-personal Kantian can accept: (1) We can inform others that what they think is their duty is not in fact their duty. For example, I can inform others that I have no desire to drink coffee, implying that their duty towards me to promote my happiness in no way requires that they buy me coffee; (2) We can set ends, or reorder our ends. Since others should promote our happiness primarily by assisting us in our own endeavours, this changes what others should do. For example, if I initially had the desire to have coffee, but noticed that I did not bring any money and would have to rely on a favour from a colleague, I might reconsider whether I really want any coffee given the circumstances. In both situations, it would be perfectly reasonable for me to tell my colleague “it’s alright, you don’t have to buy me coffee”. With my utterance I would not have lifted any duty from my colleague’s shoulders. My colleague’s duty is still what it was, namely to promote my happiness. But by signalling the conditions of my happiness, I would have provided a clue regarding the content of the other person’s duty. That is enough.

What the first-personal Kantian cannot account for, of course, is the supposed act of literally erasing someone else’s duty. Here, the ‘deny’ part of the deflate-and-deny tactic comes into play: The first-personal Kantian simply denies that this act is possible. If the obligee really has a duty towards us, it is not in our power to erase it. This has implications for the limits of our power to ‘release’ others from their duties: We can give false or misleading information about what another’s duty is (say, out of politeness). Others should not take such utterances at face value. For example, suppose I politely tell my colleague “it’s okay, you don’t have to buy me coffee”, but really I would love to have a cup. In this case, my colleague’s duty to promote my happiness still demands buying me coffee. What matters is not any internal or external act of second-personal address, but my happiness.Footnote 11

Put in this way, the first-personal Kantian view is likely much stricter than Kant’s own view. Kant evidently thinks that if X has a duty towards Y, then Y can always release X from that duty at will. This is the assumption that raises the basic problem with duties to self in the Doctrine of Virtue (MM 6:417.07–22). The problem is this: It would seem, at first sight, that duties towards self could not be stable, because we could release ourselves from these duties at will (ibid.). Evidently, this problem only arises if we understand release-from-duty as a literal erasure of duty by an act of the obligee’s will. Kant’s solution is that duties to self are really duties of the homo phaenomenon—the human being as a subject of inclinations—to the homo noumenon—the human being as a subject of a rational will (MM 6:418.14–23). But it is not as though Kant really struggled with this problem. Rather, posing and resolving the problem serves a heuristic function. Kant’s discussion illuminates the nature of duties to self: When we owe something to ‘ourselves’, we owe it to ourselves purely as rational beings. Kant could have made the same point without using this specific problem as a heuristic device. What is more, the underlying assumption that obligees can eradicate duties at will does not seem to carry any weight anywhere else in Kant’s oeuvre. Kant does not mention it in other contexts, let alone build any elaborate arguments upon it. So Kantians can confidently dispense with the assumption that obligees can eradicate duties at will. No serious difficulties follow.

What happens if we apply the first-personal view of release-from-duty to the more specific case of giving consent? On Darwall’s conception, if Y consents to X’s doing A, then Y releases X from the duty towards Y not to do A (Darwall 2012, 345). For instance, a patient may consent to a surgical procedure, thus releasing the medical personnel from the duty to refrain from that procedure, say, out of respect for bodily integrity. Darwall emphasises that by giving consent, we do more than merely to inform another that we “do not mind” (Darwall 2012, 344). On the first-personal Kantian view, however, it is just that. Or at any rate, to give consent is either to inform others about their duty or to make up one’s mind so that it becomes clear what their duty is. In certain circumstances, this may indeed amount to no more than saying that one does not mind. For instance, if a medical patient has her priorities in clear order and knows the impact of a procedure on her happiness, she can inform the medical personnel that their duty to promote her happiness does not require that they refrain from this procedure. In other cases, however, the story may be much more complicated. Perhaps the patient does not initially have a clear picture of whether she prefers the consequences of the procedure to the consequences of its omission. In such cases, obtaining the patient’s consent also forces her to make up her mind. In either case, the patient is not eradicating any duties on the part of the medical personnel, but merely makes it clearer what the content of their duties is.

The first-personal view on giving consent may invite the worry that it confers the power to consent on too many people. It seems to confer this power on anyone who can provide good information about that person’s happiness. But it does not. First, the patient’s happiness may depend on the very right to determine herself whether she allows the procedure. Human beings usually do not like being paternalised, and our practical benevolence should be sensitive to this fact.Footnote 12 Secondly, the patient of course has better epistemic access to the conditions of her own happiness and to the ends by means of which she aims to achieve it. In typical cases, patients will therefore be in a uniquely privileged position to give consent to actions which impact them. And this unique position should be our reference point when we conceive of the position in which one should be to validly give consent at all. On this conception, usually only the impacted subject will be in a position to consent.

However, exceptions are possible. There might be uninformed or misinformed obligees, and the first-personal Kantian can easily explain why we should not take their consent at face value. They are, at least in this instance, not good enough informants about the conditions of their own happiness. There could also be irrational patients: They know the ends whose realisation would make them happy, and they know the procedure is not conducive to these ends, yet they fail to rationally combine these pieces of information. Even though withholding consent would be in their best interest, they consent. Suppose there are others who have, let us assume, heard a comprehensive and true firsthand account of the patient’s desires and ends, and of the procedure’s harmful impact. On a first-personal Kantian view, those rational and informed others may be in a better position to give or deny consent than the patient. The crucial and difficult question is, of course, who is to count as uninformed, misinformed, or irrational. It merits philosophical attention in its own right. What is clear from a Kantian point of view, however, is that arrogance is a vice, and self-serving rationalisations are a straightforward danger. So we should approach the question whose consent counts with humility, even though we do not get to simply avoid it. We must not simply presume that we know better what is good for others than they themselves do.

So far, the first-personal Kantian take on consent may sound largely destructive. But there are also more constructive upshots, particularly when we think about consent as it applies to animals: For traditional, second-personal Kantians, the fact that animals cannot consent makes a world of difference.Footnote 13 It indicates that animals cannot have moral authority, do neither put us under obligation nor lift duties off our shoulders, and more generally are not proper objects of moral concern in their own right. Not so on the first-personal view. Here, the obtaining and giving of consent is merely a means of conveying information about desires, practical ends, and their ordering, or to make up our minds. What really matters at the end of the day is not an act of second-personal address. What matters is that the obligor gets the conditions of the obligee’s happiness right. Animals do not speak a human language. Still, we can take steps to get it right what makes them happy (for instance, paying close attention, counteracting our own biases, considering scientific evidence). The fact that animals cannot consent, be it verbally or by reactive attitude, is of little consequence. It merely shows that we should take alternative measures to ensure that we get the conditions of an animal’s happiness right, given that they cannot verbally tell us what our duty demands. Paying close attention to the animal and relying on the best information we have about its well-being can serve largely the same purpose as the obtaining of informed consent from human beings. It should be treated as equally morally important.

A similar difference emerges when we compare Darwall and the first-personal Kantian on the topic of forgiveness. On Darwall’s view, “forgiveness acknowledges the other’s responsibility for wronging one, but refrains from pressing claims or ‘holding it against’ him” (Darwall 2009, 72). By ‘pressing claims’ and ‘holding it against him’, Darwall means the exercise of second-personal moral authority. To ‘press claims’ is, among other things,Footnote 14 to create or keep in existence a duty for the wrongdoer to apologise, to blame herself, or to make amends for her wrongdoing. Darwall’s picture is that wronging someone leads to special duties of the wrongdoer to the wronged, and thus the wronged are obligees with a special authority to free the wrongdoer from these special duties (see Darwall 2012, 356).

We can account for part of what Darwall means by ‘forgiveness’ in first-personal Kantian terms: A verbal utterance of forgiveness can serve the purpose of informing someone that apologising, self-blame, or amends are not the content of her duty. For instance, X may forgive Y for having stolen an inconsequential amount of money several years ago. In the end, no harm came to X from Y’s theft, and the utterance of forgiveness is quite simply a way of communicating this fact. In verbally forgiving Y, X signals that the promotion of her happiness does not require reparations for this act. Perhaps in this example, verbal forgiving may be a somewhat vacuous ritual. But in other cases, verbal forgiving can convey crucial information about a person’s happiness. Say X neglected the infant Y, who has now grown up. The adult Y’s utterance of forgiveness for X may convey the—anything but obvious!—information that Y’s happiness is no longer impacted by this serious wrong. And even much more complicated arrangements are conceivable. For instance, forgiving may be an intrapersonal process through which someone who has been wronged actively emancipates herself from the impact of past wrongdoing. In forgiving, one performatively denies the wrongdoer any lasting impact. In this case, verbal forgiving is not merely an informative act, but also an act of self-assertion.

On the other hand, the first-personal Kantian cannot agree with everything that Darwall says about forgiveness. For a first-personal Kantian, forgiving others cannot be an act of lifting any duties off their shoulders. Just as the first-personal Kantian view limits our power to consent, it limits our power to forgive. Our utterances of forgiveness, like our utterances of consent, can convey false or misleading information, and in these instances we should not take them at face value (except perhaps as performative self-assertions).

When it comes to animals, the implications of the first-personal Kantian view on forgiveness echo what I have said about consent. Animals do not forgive, but on its own this is of little consequence. What matters is that we get the conditions of another’s happiness right, because that is what we must promote. And of course, the conditions of an animal’s happiness can be shaped by our past wrongs, and animals can ‘get over’ that influence over time. Animals who have suffered a certain type of abuse may have different needs than their conspecifics who have not. Our duties are sensitive to such individual differences in happiness. Hence, it is crucial that we pay close attention to how our actions have impacted an animal and how we can promote its happiness specifically against the backdrop of our wrongs.

Apologies are, among other things, pleas for forgiveness. Darwall conceives of apologies partly as pleas for the eradication of special duties that wronging someone creates, perhaps a duty to blame oneself or a duty to make reparations.Footnote 15 One quite intuitive upshot of this conception is that apologies must be directed only from the wrongdoer to the wronged (Darwall 2012, 346). Nobody else can apologise for this particular wrong, and one can apologise to nobody but the wronged. After all, Darwall’s view entails that only the wronged has the authority of an obligee to free the obligor from her duties (ibid.).

Again, the first-personal Kantian can deflate part of what Darwall means by ‘apologies’. The first hint comes from Darwall himself: “If a victim comes upon an unaddressed admission of guilt and expression of sincere regret in her victimiser’s diary, she has not discovered an apology” (ibid.). Of course, it is true that the verbal expression of apology must always be addressed to the wronged. But it is a separate question whether the duty to apologise is always owed to the person to whom the verbal utterance is addressed.

For Darwall, apologies are crucially more than admissions of guilt and regret. For the first-personal Kantian, they are basically that, plus perhaps an expression of resolve to act better in the future. To issue such utterances may be our duty to others. After all, it is quite plausible that apologies promote the happiness of those we have wronged. For instance, apologies often lead to forgiveness, and forgiveness is good for our mental health and well-being (see Toussaint and Webb 2005). So apologies can be a formidable way for wrongdoers to promote the happiness of the wronged, in a way no one else can. But apologies may also give the person we have wronged some pleasant reassurance that, having acknowledged the wrongness of their action, the wrongdoer is less likely to repeat it. The first-personal Kantian therefore has some foothold to argue that apologies are owed to the obligee based on the duty to promote their happiness. This also explains why a secret expression of guilt in a diary cannot serve the usual purpose of an apology, since it is not intended to promote the happiness of the wronged. It is equally clear, however, that exceptions are possible such that apologies do not promote the happiness of those we have wronged. For instance, if the wrong in question is very grave, an apology may indeed trivialise it. Sometimes, ‘sorry’ does not cut it. In such cases, it would indeed go against our duty towards others to apologise.

However, we may still have a duty towards self to apologise to those we have wronged. Our duties to self, on the first-personal Kantian conception, are about moral self-perfection. Apologies serve this end by helping us cultivate a sense of duty. The practice of apologising forces us to recall our wrongs and to indulge in feelings of guilt. If we additionally conceive of apologies as expressions of resolve to act better in the future, it is even more evident how this ritual promotes moral self-perfection. Apologies help us act better next time around.

The first-personal Kantian still denies that apologies should be understood as pleas to be freed from the special duties we incur by wronging others, since that plea would be impossible for others to fulfil. Obligees cannot eradicate our duties. Yet again, the focus is shifted away from the act of second-personal address and onto the purpose it serves. The apologies we ought to utter, in one way or another, serve to promote the happiness of others or our own moral perfection. What matters is that we achieve these obligatory ends rather than that we use the specific means of a verbal act of second-personal address.

Again, there are constructive upshots regarding animals. Like consent and forgiveness, apologies are means to an end. But there may be other, much more important ways to achieve that same end. This is so even with apologies when it comes to animals. Animals cannot understand apologies. Instead, we can take extra care in determining how to promote an animal’s happiness in the special context of our past wrongs. When I have stepped on a cat’s tail, I can pay special attention to whether any lasting injury has come from my wrong.

However, since we also have a duty to self to apologise, as a sort of moral exercise, it may even be our duty to issue a verbal apology to animals. As silly as it may sound at first, apologising to animals is already a common practice. Imagine a person who trips over a black cat in the dark and feels guilty for not having been more careful. It would be a perfectly understandable and reasonable reaction to exclaim “oh no, I’m sorry!” and then check if the cat is hurt and attempt to comfort her. The former is a verbal apology addressed to the animal and the latter is an alternative measure to make up for the wrong in terms of the happiness of the animal. On a first-personal Kantian view, our practice of apologising to animals is not just an irrational spillover effect from our moral interactions with humans. It can be a perfectly fine example of the fulfilment of a duty to self to apologise, based on the duty of moral self-perfection.

Let me sum up. In this chapter I have addressed a major difficulty for Kantians in including animals as moral patients: As Kant sees it, we can only have duties towards beings who share our moral law. This is because, in the view of Kant and many Kantians, our duties only have their bipolar or ‘directed’ character due to our sharing of the moral law. This view however stands in tension with the fact that, on a traditional Kantian view, the moral law is autonomous and autochthonous. It is not in any straightforward sense ‘shared’. In response, Kantians typically try to explain how the moral law is after all shared between Kantian agents. I have suggested the opposite move: Kantians should double down on the commitment that the moral law is not shared in any ambitious sense, but rather arises from the individual rational being alone. Instead of trying to account for moral bipolarity in terms of a shared moral law, we should account for it in terms of the content of our duties. Duties towards X are duties that are about X in a certain way. I have spelled out this ‘content approach’ for duties of virtue: Duties to self are about moral self-perfection, and duties to others are about others’ happiness.

Some would object that this approach is blind to an essential part of moral life, namely the part that involves second-personal authority. How could this approach explain, say, an obligee’s unique standing to give consent, to forgive a wrongdoer, or to demand apologies? Do we not need to appeal to the obligee’s capacity to put us under obligation and free us from obligation? To deal with such examples, I employ the ‘deflate-and-deny tactic’: Deflate any given part of moral life to something for which you can account in terms of first-personal Kantianism, and boldly deny the rest. I have illustrated the tactic on the three examples of consent, forgiveness, and apologies.

I hope to have shown in this chapter what first-personal Kantianism is, and that it can bring us a crucial step closer to including animals in Kantian moral concern. What remains to be done now is to see how much of Kant’s framework we can actually ‘translate’ onto our treatment of animals if we take the steps proposed so far. That is the task of the next chapter.